最近（其實(shí)也就是今天）學(xué)姐開(kāi)始了科（fan）學(xué)（qiang）上網(wǎng)，在逛github時(shí)候發(fā)現(xiàn)了一個(gè)寶藏——一位大佬分享的有關(guān)乎深度圖神經(jīng)網(wǎng)絡(luò)的相關(guān)論文。

剛好學(xué)姐最近也在整理圖神經(jīng)網(wǎng)絡(luò)的論文給微信上的小伙伴，學(xué)姐想大家伙肯定也需要這個(gè)就趕緊安排了今天的推文！

part1/經(jīng)典款論文

1. KDD 2016，Node2vec?經(jīng)典必讀第一篇，平衡同質(zhì)性和結(jié)構(gòu)性

《node2vec: Scalable Feature Learning for Networks》

2. WWW2015，LINE?1階+2階相似度

《Line: Large-scale information network embedding》

3. KDD 2016，SDNE?多層自編碼器

《Structural deep network embedding》

4. KDD 2017，metapath2vec??異構(gòu)圖網(wǎng)絡(luò)

《metapath2vec: Scalable representation learning for heterogeneous networks》

5. NIPS 2013，TransE??知識(shí)圖譜奠基

《Translating Embeddings for Modeling Multi-relational Data》

6. ICLR 2018，GAT??attention機(jī)制

《Graph Attention Network》

7. NIPS 2017，GraphSAGE??歸納式學(xué)習(xí)框架

《Inductive Representation Learning on Large Graphs 》

8. ICLR 2017，GCN?圖神經(jīng)開(kāi)山之作

《SEMI-SUPERVISED CLASSIFICATION WITH GRAPH CONVOLUTIONAL NETWORKS》

9. ICLR 2016，GGNN?門控圖神經(jīng)網(wǎng)絡(luò)

《Gated Graph Sequence Neural Networks》

10. ICML 2017，MPNN??空域卷積消息傳遞框架

《Neural Message Passing for Quantum Chemistry》

如果你不知道怎么讀論文可以先加學(xué)姐微信咱具體嘮嘮，因?yàn)樵趺醋x論文的推文下周才發(fā)，嘻嘻。

part2/熱門款論文

2020年之前

11.[arXiv 2019]Revisiting Graph Neural Networks: All We Have is Low-Pass Filters

重溫圖神經(jīng)網(wǎng)絡(luò)：我們只有低通濾波器

[論文]

https://arxiv.org/abs/1905.09550

12.[NeurIPS 2019]Break the Ceiling: Stronger Multi-scale Deep Graph Convolutional Networks

打破天花板：更強(qiáng)的多尺度深度圖卷積網(wǎng)絡(luò)

[論文]?

https://arxiv.org/abs/1906.02174

13.[ICLR 2019]?Predict then Propagate: Graph Neural Networks meet Personalized PageRank

先預(yù)測(cè)后傳播：圖神經(jīng)網(wǎng)絡(luò)滿足個(gè)性化 PageRank

[論文]?

https://arxiv.org/abs/1810.05997

[代碼]?

https://github.com/klicperajo/ppnp

14.[ICCV 2019]DeepGCNs: Can GCNs Go as Deep as CNNs?

DeepGCN：GCN能像CNN一樣深入嗎？

[論文]?

https://arxiv.org/abs/1904.03751

[代碼(Pytorch)]

https://github.com/lightaime/deep_gcns_torch

[代碼(TensorFlow)]

https://github.com/lightaime/deep_gcns

15.[ICML 2018]

Representation Learning on Graphs with Jumping Knowledge Networks

基于跳躍知識(shí)網(wǎng)絡(luò)的圖表征學(xué)習(xí)

[論文]?

https://arxiv.org/abs/1806.03536

16.[AAAI 2018]Deeper Insights into Graph Convolutional Networks for Semi-Supervised Learning

深入了解用于半監(jiān)督學(xué)習(xí)的圖卷積網(wǎng)絡(luò)

[論文]?

https://arxiv.org/abs/1801.07606

2020年

17.[arXiv 2020]Deep Graph Neural Networks with Shallow Subgraph Samplers

具有淺子圖采樣器的深圖神經(jīng)網(wǎng)絡(luò)

[論文]?

https://arxiv.org/abs/2012.01380

18.[arXiv 2020]Revisiting Graph Convolutional Network on Semi-Supervised Node Classification from an Optimization Perspective

從優(yōu)化的角度重新審視半監(jiān)督節(jié)點(diǎn)分類的圖卷積網(wǎng)絡(luò)

[論文]?

https://arxiv.org/abs/2009.11469

19.[arXiv 2020]

Tackling Over-Smoothing for General Graph Convolutional Networks

解決通用圖卷積網(wǎng)絡(luò)的過(guò)度平滑

[論文]?

https://arxiv.org/abs/2008.09864

20.[arXiv 2020]DeeperGCN: All You Need to Train Deeper GCNs

DeeperGCN：訓(xùn)練更深的 GCN 所需的一切

[論文]?

https://arxiv.org/abs/2006.07739

[代碼]

https://github.com/lightaime/deep_gcns_torch

21.[arXiv 2020]Effective Training Strategies for Deep Graph Neural Networks

深度圖神經(jīng)網(wǎng)絡(luò)的有效訓(xùn)練策略

[論文]?

https://arxiv.org/abs/2006.07107

[代碼]?

https://github.com/miafei/NodeNorm

22.[arXiv 2020]Revisiting Over-smoothing in Deep GCNs

重新審視深度GCN中的過(guò)度平滑?

[論文]?

https://arxiv.org/abs/2003.13663

23.[NeurIPS 2020]Graph Random Neural Networks for Semi-Supervised Learning on Graphs

用于圖上半監(jiān)督學(xué)習(xí)的圖隨機(jī)神經(jīng)網(wǎng)絡(luò)

[論文]?

https://proceedings.neurips.cc/paper/2020/hash/fb4c835feb0a65cc39739320d7a51c02-Abstract.html

[代碼]?

https://github.com/THUDM/GRAND

24.[NeurIPS 2020]Scattering GCN: Overcoming Oversmoothness in Graph Convolutional Networks

散射GCN：克服圖卷積網(wǎng)絡(luò)中的過(guò)度平滑

[論文]?

https://proceedings.neurips.cc/paper/2020/hash/a6b964c0bb675116a15ef1325b01ff45-Abstract.html

[代碼]?

https://github.com/dms-net/scatteringGCN

25.[NeurIPS 2020]Optimization and Generalization Analysis of Transduction through Gradient Boosting and Application to Multi-scale Graph Neural Networks

Transduction through Gradient Boosting 的優(yōu)化和泛化分析及其在多尺度圖神經(jīng)網(wǎng)絡(luò)中的應(yīng)用

[論文]?

https://proceedings.neurips.cc/paper/2020/hash/dab49080d80c724aad5ebf158d63df41-Abstract.html

[代碼]?

https://github.com/delta2323/GB-GNN

26.[NeurIPS 2020]Towards Deeper Graph Neural Networks with Differentiable Group Normalization

邁向具有可微組歸一化的更深圖神經(jīng)網(wǎng)絡(luò)

[論文]?

https://arxiv.org/abs/2006.06972

27.[ICML 2020 Workshop GRL+]A Note on Over-Smoothing for Graph Neural Networks

關(guān)于圖神經(jīng)網(wǎng)絡(luò)過(guò)度平滑的說(shuō)明

[論文]?

https://arxiv.org/abs/2006.13318

28.[ICML 2020]Bayesian Graph Neural Networks with Adaptive Connection Sampling

具有自適應(yīng)連接采樣的貝葉斯圖神經(jīng)網(wǎng)絡(luò)

[論文]?

https://arxiv.org/abs/2006.04064

29.[ICML 2020]Continuous Graph Neural Networks連續(xù)圖神經(jīng)網(wǎng)絡(luò)

[論文]?

https://arxiv.org/abs/1912.00967

30.[ICML 2020]Simple and Deep Graph Convolutional Networks簡(jiǎn)單和深度圖卷積網(wǎng)絡(luò)

[論文]?

https://arxiv.org/abs/2007.02133

[代碼]?

https://github.com/chennnM/GCNII

31.[KDD 2020]?Towards Deeper Graph Neural Networks走向更深的圖神經(jīng)網(wǎng)絡(luò)

[論文]?

https://arxiv.org/abs/2007.09296

[代碼]?

https://github.com/mengliu1998/DeeperGNN

32.[ICLR 2020]Graph Neural Networks Exponentially Lose Expressive Power for Node Classification

圖神經(jīng)網(wǎng)絡(luò)對(duì)節(jié)點(diǎn)分類的表達(dá)能力呈指數(shù)級(jí)?下降

[論文]?

https://arxiv.org/abs/1905.10947

[代碼]?

https://github.com/delta2323/gnn-asymptotics

33.[ICLR 2020]?DropEdge: Towards Deep Graph Convolutional Networks on Node Classification

DropEdge：邁向節(jié)點(diǎn)分類的深度圖卷積網(wǎng)絡(luò)

[Paper]?

https://openreview.net/forum?id=Hkx1qkrKPr

[Code]?

https://github.com/DropEdge/DropEdge

34.[ICLR 2020]?PairNorm: Tackling Oversmoothing in GNNs

PairNorm：解決GNN中的過(guò)度平滑問(wèn)題

[論文]

https://openreview.net/forum?id=rkecl1rtwB

[代碼]

https://github.com/LingxiaoShawn/PairNorm

35.[ICLR 2020]Measuring and Improving the Use of Graph Information in Graph Neural Networks

測(cè)量和改進(jìn)圖神經(jīng)網(wǎng)絡(luò)中圖信息的使用

[論文]?

https://openreview.net/forum?id=rkeIIkHKvS

[代碼]?

https://github.com/yifan-h/CS-GNN

36.[AAAI 2020]Measuring and Relieving the Over-smoothing Problem for Graph Neural Networks from the Topological View

從拓?fù)浣嵌葴y(cè)量和緩解圖神經(jīng)網(wǎng)絡(luò)的過(guò)度平滑問(wèn)題

[論文]?

https://arxiv.org/abs/1909.03211

同學(xué)們是不是發(fā)現(xiàn)有些論文有代碼，有些論文沒(méi)有代碼？學(xué)姐建議學(xué)概念讀沒(méi)代碼的，然后再讀有代碼的，原因的話上周的文章有寫(xiě)，花幾分鐘看一下【學(xué)姐帶你玩AI】公眾號(hào)的——《圖像識(shí)別深度學(xué)習(xí)研究方向沒(méi)有導(dǎo)師帶該怎么學(xué)習(xí)》

part3/最新款論文

37.[arXiv 2021]Two Sides of the Same Coin: Heterophily and Oversmoothing in Graph Convolutional Neural Networks

同一枚硬幣的兩面：圖卷積神經(jīng)網(wǎng)絡(luò)中的異質(zhì)性和過(guò)度平滑

[論文]?

https://arxiv.org/abs/2102.06462v2

38.[arXiv 2021]Graph Neural Networks Inspired by Classical Iterative Algorithms

受經(jīng)典迭代算法啟發(fā)的圖神經(jīng)網(wǎng)絡(luò)

[論文]?

https://arxiv.org/abs/2103.06064

39.[ICML 2021]Training Graph Neural Networks with 1000 Layers

訓(xùn)練 1000 層圖神經(jīng)網(wǎng)絡(luò)

[論文]?

https://arxiv.org/abs/2106.07476

[代碼]

https://github.com/lightaime/deep_gcns_torch

40.[ICML 2021]?Directional Graph Networks?方向圖網(wǎng)絡(luò)

[論文]?

https://arxiv.org/abs/2010.02863

[代碼]?

https://github.com/Saro00/DGN

41.[ICLR 2021]On the Bottleneck of Graph Neural Networks and its Practical Implications

關(guān)于圖神經(jīng)網(wǎng)絡(luò)的瓶頸及其實(shí)際意義

[論文]?

https://openreview.net/forum?id=i80OPhOCVH2

[代碼]?https://github.com/tech-srl/bottleneck/

42.[ICLR 2021]?Adaptive Universal Generalized PageRank Graph Neural Network

[論文]?

https://openreview.net/forum?id=n6jl7fLxrP

[代碼]

https://github.com/jianhao2016/GPRGNN

43.[ICLR 2021]Simple Spectral Graph Convolution

簡(jiǎn)單的譜圖卷積

[論文]

https://openreview.net/forum?id=CYO5T-YjWZV

githup原文：

https://github.com/mengliu1998/awesome-deep-gnn

以上就是學(xué)姐找到的論文了，讀論文的方法可以關(guān)注【學(xué)姐帶你玩兒AI】等到下周看學(xué)姐的推文

或者是直接私聊學(xué)姐聊一下自己的想法~~

給學(xué)姐點(diǎn)個(gè)贊吧！這是我肝文章的動(dòng)力！有問(wèn)題評(píng)論區(qū)直接找學(xué)姐！